The precise nature and neural underpinnings of schizophrenia patients' declarative memory deficits remain to be determined. Improvements in their characterization, through translation of basic cognitive neuroscience memory paradigms into clinical research, hold promise as better predictors of conversion to psychosis in adolescents at risk, better predictors of functional outcome, better targets for genome-wide association studies, and better biomarkers for testing treatment effectiveness. A critical barrier to advancing our understanding of psychosis is the lack in our understanding of the relationship between the neural underpinnings of cognitive deficits and psychotic symptoms. Tamminga, Stan, and Wagner (2010) recently proposed a hippocampus- based psychosis model of dentate gyrus (DG) hypoglutamatergic dysfunction and reduced mossy fiber input to Cornu Amonis region 3 (CA3). The model predicts that schizophrenia patients show a deficit in DG-dependent pattern separation (required for forming non-overlapping memory representations) and an increase in CA3/CA1-dependent pattern completion (required for associative retrieval based on partial information). These two processes are critically important for forming and retrieving declarative episodic memories and when disturbed may result in psychotic associations. The model can be tested using a memory paradigm sensitive to pattern separation and pattern completion, combined with high-resolution functional magnetic resonance imaging (hr-fMRI) of the hippocampus. This combination was pioneered by Stark et al. (2008), who performed a continuous incidental encoding task in which subjects were shown: objects, repeats of the objects, and lures similar to the objects. They showed that a brain activation-based bias score, which expresses similarity between lure activity as compared to repeat stimulus activity, can assess a brain region's involvement in pattern completion (bias score approaching 1; lure treated as a repetition) or pattern separation (bias score approaching 0; lure treated as first presentation). This study provided the first evidence that DG/CA3 is involved in pattern separation in humans. Moreover, Stark and colleagues have shown impaired object and spatial pattern separation performance (and associated structural and physiological abnormalities) in old versus young adults and in patients with mild cognitive impairment versus controls. In the context of hr-fMRI, the blood oxygenation level dependent signal is a function of voxel size, and although methods to correct for cardiac and respiration-associated noise exist, no study has employed them in concert with hr-fMRI, in which they may be most beneficial. We propose to investigate object and spatial pattern separation and completion- associated physiological activations with hr-fMRI in 30 chronic schizophrenia patients and 30 age- and sex- matched healthy controls, with and without corrections for cardiac and respiration-associated noise, in order to characterize the contributions of hippocampal subfields to schizophrenia patients' declarative memory deficits, neuropsychological memory deficits, symptom severity, and functional outcome levels.